A Research Based On Domain Meaning Relevance Of Depth Learning

Along with the rapid development of information technology, varies of social incidents are spread abroad fast via computer and the communications network, which are inclined to form focus incidents involved many fields and distribute the related keywords in those fields. In this process, it will be starved to resolve for the focus incident trace and the information location that how to precisely quantitatively analyze the keywords’ influence on the focus incident and the promote effect to the focus incident. Focusing on the semantic relativity of context, this paper has made a lot of research of the previous work, and lucubrated the practice of the semantic relativity of the text clustering in different fields. In specific field, it has mainly studied the problems of the conjuncture trace and the focus incidents promotion of the specific event, and realized the rapid and precise location of information. The research methods are both experiment and demonstration. The experiment is analyzed based on the real data. The brief introduction is as the following:Ⅰ. This paper has confirmed keywords of six main fields according to the classified system of Wikipedia, collected the key lemmas and their attribute lemmas, and collected these keywords from Baidu search engine and twitter, based on the breadth-first method, to support the further work.Ⅱ. This paper has studied and realized the data mining technology of new words in the field based on the statistics of term frequency, and counted the frequency of lemma’s occurrences, compared the lemmas, which are above the average threshold, with the current dictionary, in order to mining the new terms in the field.Ⅲ. This paper has studied and realized the semantics involving areas of technology, based on the deep arithmetic learning, lead terms of the specific text corpus in the vocabulary word list, mapped the terms to the real number vectors according to machine learning, and counted cosine distance between the vectors to mining the focus incidents of the language material.